---
title: "Unboxing the Box Office"
author:
output: 
  flexdashboard::flex_dashboard:
    theme:
      version: 4
      bootswatch: minty5
    orientation: columns
    vertical_layout: fill
    source_code: embed
---

```{r Setup, include=FALSE}
library(flexdashboard)
pacman::p_load(tidyverse, knitr, DT, plotly, maps, scales)
df <- read_csv("movies.csv")
world_map <- map_data("world")
options(scipen = 10)
```

<style>
.chart-title { /* chart_title */
  font-size: 20px;
  }
body{ /* Normal */
  font-size: 18px;
  }
</style>


Introduction 
===

Column {.tabset data-width=600}
---

### Introduction to the Study

**What Makes a Movie Successful**

For this final presentation, I have chosen to look at a data set containing information on various movies. The data set looks at films from the year 1980, all the way to the year 2020, giving us 40+ years worth of movies. To capture the entire industry, I believed it was best to look at it as a whole, and not choose a specific genre or rating. The movie's data set also has different writers, directors, main actors/actresses, budgets, countries of origin, and run times. With all of these variables, my goal was to find trends in the keys to "success" within films. For the sake of this study, “success” is defined solely based on ratings and gross revenue. As someone who really enjoys the film industry, I wanted to answer some of the questions I had about the correlation of certain factors that go into a movie.

My final project attempts to answer some of the questions below:

- Is there a correlation between budget and gross revenue?

- Do the ratings of a movie (good or bad) indicate the gross revenue (low or high)?

- Does the run-time of the movie affect the score?

- Do certain genres of films have a larger gross revenue than others? What about rating?

- If one were trying to create the most "successful" movie, what combination of factors have the highest odds of "success". 

### Table of Data

```{r Table}
DT::datatable(df)
```

Column {data-width=400}
---

### Variable Explanation

- **name**: This is the title of the film/movie.
- **rating**: The rating of the movie (Approved, G, PG, PG-13, R, X, Unrated, TV-PG, TV-14, TV-MA, and NC-17).
- **genre**: Genre of the movie (Action, Adventure, Animation, Biography, Comedy, Crime, Drama, Family, Fantasy, History, Horror, Music, Musical, Mystery, Romance, Sci-Fi, Sport, Thriller, and Western).
- **year**: The year of the release.
- **released**: The release date (YYYY-MM-DD).
- **score**: IMDb user rating.
- **votes**: Number of user votes.
- **director**: Name of the director.
- **writer**: Name of the writer.
- **star**: The name of the main actor/actress.
- **country**: Origin of the movie.
- **budget**: The budget of the movie in USD (Some movies don't have this, so it appears as 0).
- **gross**: Gross revenue of the movie in USD.
- **company**: The name of the production company.
- **runtime**: The length of the movie in minutes.


Rating
===

Column {.tabset data-width=600}
---

### Rating vs Revenue (1980-1989)

``` {r RvR1}
df <- df %>% 
  mutate(year = as.numeric(year))

df <- mutate(df, decade = case_when(
  year >= 1980 & year <=1989 ~ "1980's",
  year >= 1990 & year <=1999 ~ "1990's",
  year >= 2000 & year <=2009 ~ "2000's",
  year >= 2010 & year <=2020 ~ "2010's (including 2020)"))

df1980 <- df[df$decade == "1980's",]

ggplot(df1980, aes(x = score, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Score: ", score)), 
             color = "#C8A2C8", shape = 21) +
  geom_smooth(se = F, color = "#3eb489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by Rating (19980's)",
       x = "Score (out of ten)",
       y = "Gross Revenue ($)") -> p1

ggplotly(p1, tooltip = "text")
```

### Rating vs Revenue (1990-1999)

``` {r RvR2}
df1990 <- df[df$decade == "1990's",]

ggplot(df1990, aes(x = score, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Score: ", score)),
             color = "#C8A2C8", shape = 21) +
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by Rating (1990's)",
       x = "Score (out of ten)",
       y = "Gross Revenue ($)") -> p2

ggplotly(p2, tooltip = "text")
```

### Rating vs Revenue (2000-2009)

``` {r RvR3}
df2000 <- df[df$decade == "2000's",]

ggplot(df2000, aes(x = score, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Score: ", score)),
             color = "#C8A2C8", shape = 21) +
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by Rating (2000's)",
       x = "Score (out of ten)",
       y = "Gross Revenue ($)") -> p3

ggplotly(p3, tooltip = "text")
```

### Rating vs Revenue (2010-2020)

``` {r RvR4}
df2010 <- df[df$decade == "2010's (including 2020)",]

ggplot(df2010, aes(x = score, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Score: ", score)),
             color = "#C8A2C8", shape = 21)+
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by Rating (2010's)",
       x = "Score (out of ten)",
       y = "Gross Revenue ($)") -> p4

ggplotly(p4, tooltip = "text")
```

Column {data-width=400}
---

### Analysis

While looking at all scatter plots, you can see that the trend line is relatively flat across all four decades. The only time the trend line strays away from its usual shape is as we approach a score of 8 on the x-axis. As we approach the higher scores, the scatter plot reveals that gross revenue and rating have a positive correlation, whereas before this threshold there was little to no correlation at all. Out of all of the scatter plots, the 2010 decade plot has that positive correlation starting a little sooner than the others. I believe that in the future the correlation coefficient will move closer and closer to one. I theorize that in the 2010 plot, the positive correlation starts happening sooner than in the other plots due to the media. Since access to the media has drastically improved over recent years due to technological advancements, it is a lot easier to see what a movie is rated, which might affect if a viewer would want to see it more or less after seeing its score.
  


Budget
===

Column {.tabset data-width=600}
---

### Budget vs Revenue (1980-1989)

``` {r BvR1}
ggplot(df1980, aes(x = budget, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Budget: ", budget, "\n",
                              "Score: ", score)),
             color = "#613613", shape = 1) +
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Budget vs Revenue (1980's)",
       x = "Budget($)", 
       y = "Gross Revenue ($)") -> p5

ggplotly(p5, tooltip = "text")
```

### Budget vs Revenue (1990-1999)

``` {r BvR2}
ggplot(df1990, aes(x = budget, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Budget: ", budget, "\n",
                              "Score: ", score)),
             color = "#613613", shape = 1) +
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by Budget (1990's)",
       x = "Budget ($)",
       y = "Gross Revenue ($)") -> p6

ggplotly(p6, tooltip = "text")
```

### Budget vs Revenue (2000-2009)

``` {r BvR3}
ggplot(df2000, aes(x = budget, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Budget: ", budget, "\n",
                              "Score: ", score)),
             color = "#613613", shape = 1) +
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by Budget (2000's)",
       x = "Budget ($)",
       y = "Gross Revenue ($)") -> p7

ggplotly(p7, tooltip = "text")
```

### Budget vs Revenue (2010-2020)

``` {r BvR4}
ggplot(df2010, aes(x = budget, y = gross)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Revenue: ", gross, "\n",
                              "Budget: ", budget, "\n",
                              "Score: ", score)),
             color = "#613613", shape = 1) +
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by Budget (2010's)",
       x = "Budget ($)",
       y = "Gross Revenue ($)") -> p8

ggplotly(p8, tooltip = "text")
```

Column {data-width=400}
---

### Analysis

As shown previously, these plots are also broken up by decade. However, when looking at the correlation between gross revenue and budget, there seems to be an obvious positive correlation. Most of the trend lines are relatively linear. Towards a larger budget, there seems to be more of a positive correlation to gross revenue. The interesting thing to notice is the slope of the trend lines in each decade. As we get closer to the present day, the slope of the trend line gets steeper indicating that the correlation between gross revenue and budget has been becoming more positively correlated over the years.



Runtime 
===

Column {data-width=600}
---

### Runtime vs. Score

``` {r RvS1}
ggplot(df, aes(x = score, y = runtime)) +
  geom_point(aes(text = paste0(name, " \n",
                              "Runtime: ", runtime, "\n",
                              "Score: ", score, "\n",
                              "Revenue: ", gross)),
             color = "#4a6274", shape = 5) +
  geom_smooth(se = F, color = "#3EB489") +
  theme(axis.text = element_text(size = 11)) +
  labs(title = "Runtime vs Score",
       x = "Score (out of ten)", 
       y = "Runtime (minutes)") -> p9

ggplotly(p9, tooltip = "text")

summary(df$runtime)
```

Column {data-width=400}
---

### Analysis

There is little correlation between score and the runtime of a movie. Majority of the movies fall between 95 and 116 minutes (1 hour 35 minutes to 1 hour and 56 minutes), with an average run time of 107 minutes (an hour and 47 minutes). Movies that fall between these runtimes, have a variety of scores. Some are scored a little over two, while movies with the same runtime score over an eight. However, it does seem that movies that are a little longer generally score a little higher than the mean as shown by the chart's and the trend line's shape.

Genre and Rating
===

Column {.tabset data-width=600}
---

### Average Genre Revenue 

``` {r GvR}
df$gross <- as.numeric(df$gross)

df <- mutate(df, new_genre = case_when(
  genre == "Action" ~ "Action",
  genre == "Adventure" ~ "Adventure",
  genre == "Animation" ~ "Animation",
  genre == "Biograpghy" ~ "Biography",
  genre == "Comedy" ~ "Comedy",
  genre == "Crime" ~ "Crime",
  genre == "Drama" ~ "Drama",
  genre == "Horror" ~ "Horror",
  genre == "Family" | genre == "Fantasy" | genre == "History"
  | genre == "Music" | genre == "Musical" | genre == "Mystery"
  | genre == "Romance" | genre == "Sci-Fi" | genre == "Sport"
  | genre == "Triller" | genre == "Western" ~ "Other")) 

ggplot(df %>% filter(!(is.na(new_genre))), aes(x = new_genre, y = gross)) +
  geom_boxplot(color = "black", fill = "#f88379") +
  ylim(c(0,100000000)) +
  theme(axis.text = element_text(size = 11),
        axis.text.x = element_text(hjust = .6, angle = 20)) +
  labs(title = "Revenue by Genre", x = "Genre", y = "Gross Revenue ($)")
```

### Average Rating Revenue

``` {r RvR}
df <- mutate(df, new_rating = case_when(
  rating == "G" ~ "G",
  rating == "PG" ~ "PG",
  rating == "PG-13" ~ "PG-13",
  rating == "R" ~ "R",
  rating == "Approved" | rating == "NC-17" | rating == "Not Rated" |
    rating == "TV-14" | rating == "TV-MA" | rating == "TV-PG" |
    rating == "Unrated" | rating == "X" ~ "Other"))

order <- c("G", "PG", "PG-13", "R", "Other")
df$new_rating <- factor(df$new_rating, order)

ggplot(df %>% filter(!(is.na(new_rating))), aes(x = new_rating, y = gross)) +
  geom_boxplot(color = "black", fill = "#f88379") +
  theme(axis.text = element_text(size = 11)) +
  ylim(c(0,1e+08)) +
  labs(title = "Revenue by Rating",
       y = "Gross Revenue ($)",
       x = "Rating")
```

Column {data-width=400}
---

### Genre Analysis

To my personal surprise, majority of the genres have a similar median. The three highest medians are Action, Animation and Horror. However, animation has the largest third quartile. I was surprised to see that animation and horror had some of the highest medians.

### Rating Analysis

The box plots by rating didn't come as too much as a surprise since majority of movies are PG-13. This could also partial explain why their third and first quartile are respectively higher and lower to the other genres.

Other Factors
===

Column {.tabset data-width=600}
---

### Company

``` {r Company}
df_company0 <- df %>% 
  filter(gross > mean(gross, na.rm = T)) %>% 
  filter(score > mean(score, na.rm = T))

df_company1 <- df_company0 %>% 
  group_by(company) %>%
  summarise(n = n()) %>%
  arrange(desc(n)) %>%
  .[1:10,] 

df_company <- df_company0 %>% 
  semi_join(df_company1, by = "company")
  
ggplot(df_company, aes(y = gross, x = company)) +
  geom_boxplot(color = "#7851A9", fill = "darkgrey") +
  ylim(c(0,1700000000)) +
  theme(axis.text.x = element_text(angle=20, hjust=.8),
        axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by the Top 10 Companies",
       x = "Companies",
       y = "Gross Revenue ($)")
```

### Writer

``` {r Writer}
df_writer0 <- df %>% 
  filter(gross > mean(gross, na.rm = T)) %>% 
  filter(score > mean(score, na.rm = T)) 

df_writer1 <- df_writer0 %>% 
  group_by(writer) %>%
  summarise(n = n()) %>%
  arrange(desc(n)) %>%
  .[1:10,] 

df_writer <- df_writer0 %>% 
  semi_join(df_writer1, by = "writer")

ggplot(df_writer, aes(y = gross, x = writer)) +
  geom_boxplot(color = "#7851A9", fill = "darkgrey") +
  ylim(c(0,1250000000)) +
  theme(axis.text.x = element_text(angle=20, hjust=.8),
        axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by the Top 10 Writers",
       x = "Writers",
       y = "Gross Revenue ($)")
```

### Director

``` {r Director}
df_director0 <- df %>% 
  filter(gross > mean(gross, na.rm = T)) %>% 
  filter(score > mean(score, na.rm = T)) 

df_director1 <- df_director0 %>% 
  group_by(director) %>%
  summarise(n = n()) %>%
  arrange(desc(n)) %>%
  .[1:10,] 

df_director <- df_director0 %>% 
  semi_join(df_director1, by = "director")

ggplot(df_director, aes(y = gross, x = director)) +
  geom_boxplot(color = "#7851A9", fill = "darkgrey") +
  theme(axis.text.x = element_text(angle=20, hjust=.8),
        axis.text = element_text(size = 11)) +
  labs(title = "Gross Revenue by the Top 10 Directors",
       x = "Director",
       y = "Gross Revenue ($)")
```

### Star Actor/Actress

``` {r Star}
df_star0 <- df %>% 
  filter(gross > mean(gross, na.rm = T)) %>% 
  filter(score > mean(score, na.rm = T))

df_star1 <- df_star0 %>% 
  group_by(star) %>%
  summarise(n = n()) %>%
  arrange(desc(n)) %>%
  .[1:10,] 

df_star <- df_star0 %>% 
  semi_join(df_star1, by = "star")

ggplot(df_star, aes(x = star, y = gross)) +
  geom_boxplot(color = "#7851A9", fill = "darkgrey") +
  theme(axis.text.x = element_text(angle=20, hjust=.8),
        axis.text = element_text(size = 11)) +
  ylim(c(78500541, 1000000000)) +
  labs(title = "Gross Revenue by the Top 10 Actors/Actresses",
       x = "Actor/Actress",
       y = "Gross Revenue ($)") 
```

Column {data-width=400}
---

### Analysis and Conclusion

When looking at the company box plot the three highest medians are DreamWorks Animation, Walt Disney Pictures, and Twentieth Century Fox. The most noticeable difference between these three is the third and first quarter for Disney. They have a massive distribution.

Taking a look at the writer box plot, Fran Walsh (wrote King Kong in 2005 and wrote the Hobbit Trilogy), Christopher Markus (wrote all the Captain America movies along with the last two Avengers movies), and John Lasseter (wrote all the Toy Story and Cars movies) have the largest medians. Fran Walsh however, has the largest median by far.

Looking at the director box plot Peter Jackson (directed all the Lord of the Rings) stands out like an eyesore. Not only do they have the largest median, but have the largest first quartile by a mile. After Peter, Christopher Nolan (directed the Bale Batman trilogy, Inception, and Interstellar) also does incredibly well.

Lastly, the actors and actresses. Most of them float around the same median with only Leonardo DiCaprio (starred in the Titanic, Romeo and Juliet, Inception, and the Wolf of Wall Street) and Will Smith (starred in the Bad Boys series, I am Legend, and The Pursuit of Happiness) being stand out actors in this plot.

Recalling our definition of success, to have the highest chance of a "successful" movie these are the components you should think about incorporating. You should either have an Action or Animated movie with a PG-13 rating. It would be ideal to work for DreamWorks or Disney and have Fran Walsh as your writer. To cap off your team, you pick up Peter Jackson as lead director, with Leonardo DiCaprio or Will Smith as your lead actor.



Map
===

Column {data-width=600}
---

``` {r Map}
# Rename some country names to match the names in the dataset to match the map data
df$country <- recode(df$country,
                     "United Kingdom" = "UK",
                     "United States" = "USA",
                     "West Germany" = "Germany",
                     "Soviet Union" = "Russia",
                     "Hong Kong" = "China",
                     "Yugoslavia" = "Serbia",
                     "Republic of Macedonia" = "North Macedonia",
                     "Federal Republic of Yugoslavia" = "Serbia")

# Get summary statistics for the map
df_summary <- df %>%
  group_by(country) %>%
  summarise(score = round(mean(score, na.rm = T)),
            budget = round(mean(budget, na.rm = T)),
            gross = round(mean(gross, na.rm = T)),
            runtime = round(mean(runtime, na.rm = T)),
            count = n())

# Merge the map data to the movie summary
df_map <- df_summary %>%
  left_join(world_map, by = c("country" = "region"))

# Create the map
world_map %>%
  ggplot() +
  geom_polygon(aes(x = long, y = lat, group = group,
                   text = paste0("Country: ", region)),
               fill="lightgrey") +
  geom_polygon(data = df_map,
               aes(x = long, y =lat,
                   group = group, fill = gross,
                   text = paste0("Country: ", country, "\n",
                                 count, " movies\n",
                                 "Avg. Budget: ", comma(budget), " US dollars\n",
                                 "Avg. Gross: ", comma(gross), " US dollars\n",
                                 "Avg. Score: ", comma(score), " (out of 10)\n",
                                 "Avg. Runtime: ", comma(runtime), " minutes"))) +
  scale_fill_gradientn(colors = c("#a7c0cd", "#68a0b0", "#31777d", "#004d40")) +
  theme_void() +
  theme(legend.position = "none") -> p

font <- list(
  family = "Arial",
  size = 15,
  color = "white"
)

label <- list(
  bgcolor = "#232F34",
  bordercolor = "transparent",
  font = font
)

ggplotly(p, tooltip = "text",
         width = 1200,
         height = 600) %>%
  style(hoverlabel = label) %>%
  layout(font = font)
```

Column {data-width=400}
---

### Analysis

The map shows each country's average budget, gross revenue, score, and runtime. Surprisingly, China and Finland had a higher average gross revenue than the United States. However, these countries also have a higher average budget than the United States. Not only that, but the United States only averages a score of 6 on IMBD whereas the majority of counties average a 7, with some averaging an 8!



About the Author
===

Column {data-width=500}
---

### Who am I?

My name is Jonah Mergler and I am an undergraduate student at the University of Dayton studying Applied Mathematical Economics major with a minor in Data Analytics. I am on track to graduate in May 2025.

The summer after my sophomore year and into my 5th semester at UD, I was a Financial Analysts inter with an insurance broker firm named McGohan Brabender. During my time at McGohan Brabender, I worked with the Stop Loss model for the 100+ segment and created a lot of medical renewals for a large number of unique organizations.

My goal is to acquire an internship outside of the insurance industry to see what else the math workforce has else to offer. I am actively seeking employment on both Handshake and LinkedIn. If you are interested in reaching out, my linked in profile can be found [here](https://www.linkedin.com/in/jonah-mergler-dayton). 

Column {data-width=500}
---

### Picture of Me

``` {r Picture, fig.width=6, echo=FALSE, fig.cap="My friend Sam(left) and I(right) in the streets of San Deigo, California."}
knitr::include_graphics("Sam&MeCali.jpg")
```